Abstract In this work the type XVIII collagen has been studied by using several approaches, such as different animal models. The primary structure of frog, Xenopus laevis, type XVIII collagen and the expression pattern of its variants during early embryogenesis have been elucidated. The gene structure of human type XVIII collagen was characterized and the localization and processing of its longest variant was studied by generated antibodies. In addition, the function of the proteolytically released C-terminal part of type XVIII collagen, endostatin, was studied by generating transgenic mice expressing endostatin. The primary structure of X. laevis type XVIII collagen is comprised of three N-terminal variants resembling their mammalian counterparts. The sizes of the polypeptides are 1285, 1581, and 1886 residues. The most conserved regions are the C-terminal endostatin region and the cysteine-rich domain in the N-terminus. Whole-mount in situ hybridization reveals different expression patterns for variants during embryogenesis. The short variant is the most abundant, whereas the two longest variants exhibit more restricted expression. The gene structure of human type XVIII collagen reveals an exon-intron organization that is conserved with mouse. The length of the human gene is about 105 kb and contains 43 exons. The third variant of type XVIII collagen has a conserved cysteine-rich domain with homology to the extracellular part of frizzled proteins. This third variant is localized to developing muscle and lung, and is also found in serum. In cell culture, the proteolytic fragments of the N-terminus, including the cysteine-rich motif, are also detected. Endostatin function was studied by generating mouse lines expressing endostatin under the keratin-14 promoter, which drives the expression mainly in the skin. Three independent transgenic mouse lines were achieved with varied expression levels. The phenotype was seen in the eye with lens opacity and abnormal morphology of epithelial cells in the lens. In the skin, a broading of the basement membrane in the epidermis dermis junction was detected. Immunoelectron microscopy analysis revealed a polarized orientation of type XVIII collagen in the basement membrane. In transgenic mice, altered localization of endogenous type XVIII collagen was seen, suggesting displacement of the endogenous type XVIII collagen with transgenic endostatin leading to disorganized basement membrane.

Diss. Linköping : Linköpings universitet, 2007.

Abstract Cancer progression is a complex multi-step process. Two critical steps in tumor growth and invasion are the proteolytic processing of the extracellular matrix environment and the angiogenic switch enabling blood supply into the tumor. Matrix metalloproteases (MMPs) are a group of proteolytic enzymes involved in physiological and pathological extracellular matrix processing. Trypsinogen, a serine protease, is one of the first proteolytic enzymes characterized. The amount of one of its isoforms, tumor associated trypsinogen-2 (TAT-2) correlates with the malignant phenotype of several forms of cancers. Both of these protease groups are critically dependent on their activation from latent proforms to fully active enzymes. We found that the overproduction of TAT-2 in malignant oral squamous cell carcinoma cell line was associated with elevated proMMP-9 (but not proMMP-2) activation, as well as enhanced cancer cell intravasation in an in vivo model. This indicates that TAT-2 and MMP-9 activation play a role in the invasive growth of oral carcinomas. Proteases are involved in angiogenesis, the formation of new blood vessels, in several ways. One mechanism is the release of cryptic anti-angiogenic molecules from larger extracellular matrix components. Endostatin is one such cryptic endogenous inhibitor of angiogenesis. Certain MMPs were able to cleave endostatin from its parent molecule, collagen XVIII. The endostatin fragments generated by MMP-3, -7, -9, -13 and -20 inhibited angiogenesis in a similar fashion as the native endostatin. The regulation between MMPs and endostatin was shown to be reciprocal, as endostatin was able to block the activation and activities of MMP-2, -9 and -13. The inhibition of these tumor-associated MMPs explains at least in part the anti-tumor activity of endostatin. Endostatin not only affects endothelial cell growth as is usually thought, but it also inhibits the migration of oral carcinoma cells. In addition, cell density and proper concentration were proven to be critical for the activity of endostatin. Arresten is another endogenous inhibitor of angiogenesis and tumor growth derived from type IV collagen. We confirmed that arresten binds to integrin α1β1 on endothelial cell surface. We found that this binding is functionally significant for the anti-angiogenic properties of arresten, as tumors planted to integrin α1 knockout mice or endothelial cells derived from those mice did not respond to arresten treatment.

Abstract Collagen XVIII is a basement membrane proteoglycan, which has three variant N-termini. These variants are coded by two promoters; promoter 1 directs the synthesis of a short variant and promoter 2 directs the synthesis of two longer variants, of which the middle variant is generated from the longest by splicing. The longest variant contains a cysteine-rich domain in its N-terminus, which shows homology to the frizzled receptors of the Wnt molecules and can inhibit Wnt/beta-catenin signalling in vitro. The C-terminal domain of collagen XVIII, endostatin, is an inhibitor of tumor growth and angiogenesis. Lack of collagen XVIII accelerates cutanous wound healing and wound angiogenesis. Overexpression of endostatin leads to delayed wound healing and the presence of morphologically abnormal wound capillaries. Moreover, endostatin overexpression leads to delayed formation of the wound epidermal basement membrane and impaired maturation of hemidesmosomes. Endostatin treatment decreases osteoblast proliferation in vitro. Moreover, osteoblast proliferation and mineralization of the matrix by osteoblasts are inhibited when cells are treated with endostatin together with VEGF. In vivo, lack of collagen XVIII leads to delayed formation of secondary ossification centers in mouse femurs, whereas overexpression of endostatin leads to a slower growth of bone length. However, both of these changes are transient and mild, suggesting that collagen XVIII/endostatin is not essential for skeletal development. The growth of hair follicles is delayed in the mice overexpressing endostatin. This delay in growth is preceded by an impaired hair follicle associated angiogenesis. Lack of collagen XVIII causes an accelerated onset of the first hair cycle. A similar change can be seen in mice lacking the long variants of collagen XVIII. Lack of the short variant causes mild acceleration in the catagen of the first cycle, and anagen is also significantly accelerated in these mice. The long variants were located in the bulge region, which contains the hair follicle stem cells, and in the basement membrane surrounding the dermal papilla. As it is known that several Wnt-inhibitors are upregulated in the bulge, our results suggest that the longest variant of collagen XVIII may have a role as a regulator of Wnt-signalling in hair follicles.